General Report on Tunny
25C Page 160
For each run the following particulars are entered (horizontally: see exhibits).
Number of run (corresponding to that on the Colossus run sheet)
Number of messages used
Wheels used (the various incomplete wheels are named systematically, ΔΧ1A, ΔΧ1B, ΔΧ1C e.g. on the chi 4 sheet. BB--A means that ΔΧ1B, ΔΧ2B, ΔΧ3A, were used, ΔΧ2 being as yet unknown).
Spanning, limitation, doubting, etc. (if needed).
The run used.
Decibans per pip, and/or a statement (simply PIPS) that results are entered in pips.
The score for each character (in pips or decibans).
 |
generally omitted when the wheels are complete enough to use | |
| letter counts for decibanning. | ||
Other particulars are recorded in a log book. (cf R4 p 19.)
(c) Choosing wheel-breaking runs.
In setting, the break-in is usually followed by another two-wheel run. In breaking, the rectangle is NOT usually followed by another two-wheel run, i.e. a conditional rectangle, because of the time required: the neglect of conditional rectangles has perhaps been excessive. (See 24F.)
It is very easy to see that the conditions for the success of a given short run in setting and breaking are similar but not identical, for the criterion in setting is sigma-age
The sigma-age of a wheel-breaking run on its wheel is
Significance depends on
Decibanage per pip depends on
The average number of pips per character is
Evidently, when there is a choice of the wheel to be run for, a weaker run for a shorter wheel may be preferable.
Because of different decibannages per pip, and the possibility that one of the component runs may be very weak there is rather more advantage than when setting in keeping run separate e.g. 5.1.2. and 5x1x2x rather than 5=1=2. If the two runs can profitably be added this is evident to the eye. (R2 p 14, but see R2 p 62.)
Back to General Report on Tunny. Contents.